ABSTRACT This proposed study will explore the relationship between reported tobacco use and presence/amount of cotinine (an objective biomarker of smoking) in existing specimens linked to a large, widely available dataset often used for epidemiologic research, including for childhood cancer studies. Tobacco constituents cross the placenta, and may place pregnancies at risk of adverse outcomes including childhood cancer. Thus, epidemiologic studies of cancer etiology routinely obtain data about smoking, virtually always by self-report. Individuals may inaccurately report smoker status due to social desirability, or because data are collected many years after diagnosis. Birth records are an important data source for studies of many outcomes, and linked cancer registry-birth records are increasingly being used in collaborative case-control studies of childhood cancer. One advantage is that prenatal smoking information on birth records is obtained prior to disease diagnosis, and near the pregnancy. We propose to assess the accuracy of self-reported maternal prenatal smoking information in the birth record, by using newly developed laboratory methods in combination with dried blood spots (DBS) from newborn screening archives. We propose to measure the cotinine in DBS specimens for 200 infants identified in Washington State birth records from 2008: 100 born to randomly selected women who indicated they smoked during pregnancy, and 100 born to non-smokers (25 randomly selected from each of 4 different smoking dose levels). Specifically, we propose to: 1) Describe the presence and levels of cotinine among reported non-smokers and smokers, and among smokers by dose (1-9, 10-19, 10-19, 30+ cigarettes/day) and duration (part vs. all of pregnancy), and 2) Assess the validity of self-reported prenatal smoking status (yes/no) by comparing this information to cotinine levels in DBS. This information will be important for future childhood cancer studies, which increasingly are accessing archived DBS as a source of DNA for gene studies, and for which DBS may also provide biomarkers of environmental exposures. PUBLIC HEALTH RELEVANCE: PROJECT NARRATIVE: Assessing smoking as a risk factor in epidemiology studies of childhood cancer is difficult. We propose to measure the accuracy of self-reported prenatal smoking information by comparing levels of cotinine (a biomarker of cigarette smoking) in dried blood spots of 200 infants born to reported smokers and non-smokers identified in birth records. This information may help improve studies of childhood cancer and other outcomes such as malformations and Sudden Infant Death Syndrome.